半导体各工艺简介-2

88die200-mmwafer232die300-mmwaferIncreaseinNumberofChipsIncreaseinNumberofChipsonLargerWaferDiametersonLargerWaferDiameters(Assumelarge1.5x1.5cmmicroprocessors)(Assumelarge1.5x1.5cmmicroprocessors)WaferPolishingDoubleDouble--SidedWaferPolishSidedWaferPolishUpperpolishingpadLowerpolishingpadWaferSlurryFigure4.26ImprovingSiliconWaferRequirementsImprovingSiliconWaferRequirementsYear(CriticalDimension)1995(0.35µm)1998(0.25µm)2000(0.18µm)2004(0.13µm)Waferdiameter(mm)200200300300SiteflatnessA(µm)Sitesize(mmxmm)0.23(22x22)0.17(26x32)0.1226x320.0826x36MicroroughnessBoffrontsurface(RMS)C(nm)0.20.150.10.1Oxygencontent(ppm)D≤24±2≤23±2≤23±1.5≤22±1.5BulkmicrodefectsE(defects/cm2)≤5000≤1000≤500≤100Particlesperunitarea(#/cm2)0.170.130.0750.055EpilayerFthickness(±%uniformity)(µm)3.0(±5%)2.0(±3%)1.4(±2%)1.0(±2%)AdaptedfromK.M.Kim,“BiggerandBetterCZSiliconCrystals,”SolidStateTechnology(November1996),p.71.硅单晶片的制备：单晶切割分段结晶性与电阻率检测清洗腐蚀滚磨外圆定位面研磨切片清洗磨片清洗抛光清洗检验QualityMeasuresQualityMeasures„„PhysicaldimensionsPhysicaldimensions„„FlatnessFlatness„„MicroroughnessMicroroughness„„OxygencontentOxygencontent„„CrystaldefectsCrystaldefects„„ParticlesParticles„„BulkresistivityBulkresistivityBasicprincipleofthehotprobe,illustratedforanBasicprincipleofthehotprobe,illustratedforanNN--typesample,fordeterminingNtypesample,fordeterminingN--orPorP--typetypebehaviorinsemiconductors.behaviorinsemiconductors.HotPointProbeHotPointProbeConsideranNConsideranN--typesample.Themajoritycarriersareelectrons.Atthetypesample.Themajoritycarriersareelectrons.Atthehotprobe,thethermalenergyoftheelectronsishigherthanathotprobe,thethermalenergyoftheelectronsishigherthanatthecoldthecoldprobesotheelectronswilltendtodiffuseawayfromthehotprprobesotheelectronswilltendtodiffuseawayfromthehotprobe,obe,drivenbythetemperaturegradient.Ifawirewereconnectedbetdrivenbythetemperaturegradient.Ifawirewereconnectedbetweenweenthehotandcoldprobes,thiswouldresultinameasurablecurrethehotandcoldprobes,thiswouldresultinameasurablecurrent,nt,whosedirectionwouldcorrespondtotheelectronsmovingrighttwhosedirectionwouldcorrespondtotheelectronsmovingrighttoleft.oleft.(Thecurrentbydefinitionwouldbeintheoppositedirection.)(Thecurrentbydefinitionwouldbeintheoppositedirection.)IfweIfweplaceahighplaceahigh--impedancevoltmeterbetweentheprobes,nocurrentflows,impedancevoltmeterbetweentheprobes,nocurrentflows,butapotentialdifferenceismeasured,asillustrated.Astheebutapotentialdifferenceismeasured,asillustrated.Astheelectronslectronsdiffuseawayfromthehotprobe,theyleavebehindthepositiveldiffuseawayfromthehotprobe,theyleavebehindthepositivelyycharged,immobiledonoratomsthatprovidedtheelectrons.Thecharged,immobiledonoratomsthatprovidedtheelectrons.Thenegativelychargedmobileelectronstendtobuildupneartheconegativelychargedmobileelectronstendtobuildupnearthecoldprobe.ldprobe.ThisresultsinthehotprobebecomingpositivewithrespecttoThisresultsinthehotprobebecomingpositivewithrespecttothecoldthecoldprobe.Byasimilarsetofarguments,ifthematerialwerePtypprobe.Byasimilarsetofarguments,ifthematerialwerePtype,e,positivelychargedholeswouldbethemajoritycarriersandthepositivelychargedholeswouldbethemajoritycarriersandthepolaritypolarityoftheinducedvoltagewouldbereversed.Thedirectionofthecoftheinducedvoltagewouldbereversed.ThedirectionofthecurrenturrentbetweenthetwoprobeswouldalsobereversedinPbetweenthetwoprobeswouldalsobereversedinP--typematerial,iftypematerial,iftheywereshortedwithawire.Thusameasurementofeitherthetheywereshortedwithawire.Thusameasurementofeithertheshortshort--circuitcurrentortheopencircuitvoltagetellsusthetypeofcircuitcurrentortheopencircuitvoltagetellsusthetypeofthematerial.thematerial.““FourFour--pointprobepointprobe””measurementmethod.Themeasurementmethod.Theoutertwoprobesforceacurrentthroughtheoutertwoprobesforceacurrentthroughthesample;theinnertwoprobesmeasurethevoltagesample;theinnertwoprobesmeasurethevoltagedrop.drop.MeasurementofSheetResistanceMeasurementofSheetResistance„„ThemostcommonmethodofmeasuringthewaferresistivityisThemostcommonmethodofmeasuringthewaferresistivityiswiththewiththefourfour--pointprobepointprobe.Wemeasurethesample.Wemeasurethesampleresistancebymeasuringthecurrentthatflowsforagivenappliresistancebymeasuringthecurrentthatflowsforagivenappliededvoltage.Thiscouldbedonewithjusttwoprobes.However,involtage.Thiscouldbedonewithjusttwoprobes.However,inthatcase,contactresistancesassociatedwiththeprobesandthatcase,contactresistancesassociatedwiththeprobesandcurrentspreadingproblemsaroundtheprobesareimportantandcurrentspreadingproblemsaroundtheprobesareimportantandarenoteasilyaccountedforintheanalysis.Usingfourprobesarenoteasilyaccountedforintheanalysis.Usingfourprobesallowsustoforcethecurrentthroughthetwoouterprobes,allowsustoforcethecurrentthroughthetwoouterprobes,wheretherewillstillbecontactresistanceandcurrentspreadiwheretherewillstillbecontactresistanceandcurrentspreadingngproblems,butwemeasurethevoltagedropwiththetwoinnerproblems,butwemeasurethevoltagedropwiththetwoinnerprobesusingahighpr